Comprehensive, up-to-date information on HIV/AIDS treatment, prevention, and policy from the University of California San Francisco

Bacterial infections have emerged as an important cause of morbidity and mortality in individuals infected with the human immunodeficiency virus (HIV-1). Persons with HIV-1 infection are more susceptible to bacterial infections because of defects in both cell-mediated and humoral immunity. HIV-1 seropositive individuals are particularly susceptible to infections with encapsulated bacteria, such as Streptococcus pneumoniae. In fact, pneumococcus is one of the most common bacterial pathogens affecting both HIV-1 infected children and adults.(1,2) Among female sex workers in Nairobi, Kenya, S. pneumoniae caused more disease than Mycobacterium tuberculosis or non-typhi salmonellae.(3) Although many different sites of pneumococcal infection have been described,(4) the respiratory tract is the primary focus of disease.(5) Fortunately, the management of pneumococcal disease is the same regardless of the patient's HIV-1 serostatus, although the optimal preventive strategy has yet to be determined. As antimicrobial resistance increases in S. pneumoniae, it will become even more important to develop effective preventive strategies. This chapter reviews the epidemiology, pathogenesis, clinical presentation, and management of infections caused by S. pneumoniae in HIV-1 infected patients.

Population-based studies have demonstrated that the incidence of pneumococcal disease in persons with HIV-1 infection is extremely high. In San Francisco, the estimated rate of pneumococcal bacteremia in AIDS patients was 9.4 cases per 100 person-years,(6) over 150-fold higher than an age-matched population.(7) The relative risk of pneumococcal bacteremia among persons with HIV-1 infection in Franklin County, Ohio was 42 times that of county residents 18 to 64 years of age.(8) Among AIDS patients in New Jersey, 25 to 44 years of age, the incidence of invasive pneumococcal disease was 1.07 per cases per 100 persons per year compared with 0.003 per 100 persons per year in seronegative individuals(9): the incidence of pneumococcal disease was almost 300 times greater than that in age-matched populations.

The respiratory tract is the most common site of invasive pneumococcal infection in HIV-1 infected persons. In a retrospective study from Yale,(5) HIV-1 seropositive patients were more likely than seronegative patients (92% versus 63%) to have pneumonia as the source of bacteremia. Other sources of bacteremia included otitis media (10%), meningitis (5%), sinusitis (5%), and primary bacteremia (5%). Wallace and coworkers described the pulmonary diseases in HIV-1 infected individuals in a prospective multicenter study in which subjects were followed for 18 months.(10) Overall, bacterial pneumonia (4.9%) was more common than Pneumocystis carinii pneumonia (3.9%). S. pneumonia was the most common pathogen isolated: 13 of 30 cases in which a pathogen was isolated.

The risk factors for the development of bacterial pneumonia were evaluated in a multicenter observational study in which 1,130 patients with HIV-1 infection were monitored for nearly 5 years.(11) The rate of bacterial pneumonia among the HIV-1 seropositive subjects was 5.5 per 100 person-years compared with 0.9 per 100 person-years in the seronegative controls. The risk of pneumonia increased as the CD4 lymphocyte count declined and the rate was highest in those persons with CD4 counts less than 200 cells/mm³. Injection drug users had a higher rate of bacterial pneumonia than did other transmission categories. In a nested case-control study, investigators at Johns Hopkins Medical Center reported that HIV-1 infected patients with pneumococcal pneumonia were more likely than controls to be African-American, have <200 CD4 cells/mm³, have a history of previous pneumonia, and have an albumin <3.0 g/dL.(12) The use of zidovudine and previous pneumococcal vaccination when the subject had >200 CD4 cells/mm³ were less common in cases than in controls.

Pneumococcal disease can occur at any time during the course of HIV-1 infection. In a retrospective study of hospitalized HIV-1 seropositive patients with pneumococcal bacteremia, the mean CD4 lymphocyte count in children was 778 cells/mm³ (range, 62-1,820) and in adults, 185 cells/mm³ (range, 0-700).(5) The mean CD4 lymphocyte count of female sex workers in Nairobi, Kenya with invasive pneumococcal disease was 302 cells/mm³.(3) Although pneumococcal disease can be an early manifestation of HIV-1 infection, its incidence increases as HIV-1 disease progresses.(10-12)

Children with advanced HIV-1 disease are also at increased risk for serious bacterial infections: S. pneumoniae accounts for at least one-third of these infections.(13,14) Among HIV-1 infected children, bacterial pneumonia occurs in 19% to 63% and is associated with bacteremia in 15% to 25%.(13,15) Primary pneumococcal bacteremia is more common in HIV-1 infected children than in HIV-1 infected adults.(13,14) In a population-based study of children with HIV-1 infection, risk factors for invasive pneumococcal disease included an AIDS diagnosis and high levels of total serum IgG and IgM compared with HIV-1 seropositive controls.(2)

Resistance of S. pneumoniae to penicillin and other antimicrobial agents is increasing in many parts of the world, including the United States. In a survey of 13 hospitals in the United States, 7% of the pneumococcal isolates were resistant to penicillin.(16) In Atlanta, 25% of 431 isolates from patients with invasive pneumococcal infections were resistant to penicillin (7% were highly resistant).(17) Twenty-six percent of the isolates were also resistant to TMP-SMX (7% were highly resistant), 15% were resistant to erythromycin, 9% to cefotaxime, and 25% to multiple drugs.(17) Children 6 years of age and younger and whites were more likely to harbor resistant organisms.

The relationship between HIV-1 infection and penicillin resistance has not been studied extensively. A nationwide retrospective study from France(18) demonstrated that age <15 years, isolation of the organism from the upper respiratory tract or from the sinuses or middle ear, HIV infection, beta-lactam antibiotic therapy in the previous 6 months, and nosocomial acquisition were all independently associated with harboring penicillin-resistant isolates. A different study,(19) also from France, revealed that 14 (31%) of 45 hospitalized HIV-1 seropositive patients with pneumococcal disease had isolates resistant to penicillin. Treatment with antibiotics, particularly TMP-SMX, in the previous 3 months was associated with decreased susceptibility to penicillin G.

HIV-1 infection is responsible for immunologic defects not only in T lymphocytes, but also in B lymphocytes, macrophages, polymorphonuclear cells, and cytokine production. Although the primary immunologic defect is in cell-mediated immunity, defects in the humoral immune system also occur and are early immunologic sequelae of HIV infection.(20) The numerous abnormalities in the humoral immune system and phagocytic cells increase the susceptibility of the HIV-1 infected person to bacterial infections, particularly those caused by encapsulated organisms.

In the setting of HIV-1 infection, the predisposition to pneumococcal infection is probably due to dysfunctional host defenses rather than to increased colonization. Rates of pharyngeal carriage with S. pneumoniae in ambulatory adults in Denver were similar among HIV-1 infected patients (14% colonized) and HIV-seronegative controls (9%).(21) One study(22) revealed that the frequency of colonization with S. pneumoniae among participants of a vaccine study was 7% in subjects with <200 cells/mm³, 20% in subjects with at least 200 cells/mm³, and 10% in controls.

Defects in mucosal immunity could play a role in the susceptibility to pneumococcal infections because mucosal IgA may prevent bacterial adherence to mucosal surfaces and colonization with S. pneumoniae.(1) In one study, levels of secretory IgA (isotypes IgA1 and IgA2) were depressed in patients with AIDS compared to persons with asymptomatic HIV-1 infection and HIV-1 seronegative healthy controls.(23) Another study, however, reported that the levels and proportions of salivary IgA2 relative to total IgA were similar among HIV-1 seropositive and seronegative subjects with pneumococcal bacteremia.(24) Thus, other defects in mucosal cellular responses and systemic immunity are likely to play a more important role in predisposing HIV-1 infected patients to invasive pneumococcal disease.

HIV-1 nonspecifically activates B lymphocytes, leading to a polyclonal hypergammaglobulinemia. While this activation results in increased circulating immunoglobulin levels, some HIV-1 infected patients are unable to produce a specific antibody response to acute infections or vaccines.(25-28) Failure to generate pneumococcal capsular-specific IgM appears to be the most significant defect in antibody production.(29) Moreover, although the levels of total IgG are often elevated after infection with HIV-1, levels of IgG2 are frequently low.(29) Because IgG2 is often a predominant subclass produced in response to pneumococcal polysaccharides, it is possible that low levels of IgG2 may predispose patients with HIV-1 infection to pneumococcal disease.(1) Complement is important in opsonizing and killing S. pneumoniae. In HIV-infected patients, abnormalities in complement activation(30) and macrophage-dependent clearance of opsonized particles(31) have been demonstrated, although the clinical significance of these findings is unknown.

Bacterial pneumonia, including pneumococcal pneumonia, has been shown to cause a transient increase in HIV RNA levels in AIDS patients receiving antiretroviral drugs.(32) If repeated episodes of pneumonia are demonstrated to lead to a more rapid progression of HIV disease, it will be extremely important to find an effective way to prevent pneumococcal infections.

The approach to the diagnosis of pneumococcal infections is the same, regardless of HIV-1 serostatus. For patients with community-acquired pneumonia, the sputum Gram's stain may suggest the diagnosis if there are few epithelial cells, many polymorphonuclear leukocytes, and a predominant bacterial morphology.(48) Blood cultures are very specific and clinicians should obtain them in all HIV-1 infected patients suspected of having serious pneumococcal infection. The incidence of pneumococcal bacteremia in patients with HIV-1 infection and pneumonia has ranged from 50% to more than 75%.(33,49) In comparison, in HIV-seronegative patients who have pneumococcal pneumonia and are younger than 60 years of age, the incidence of bacteremia is less than 30%.(50)

Patients suspected of having an extrapulmonary source of infection should have appropriate fluids examined with standard laboratory and microbiologic tests. Individuals suspected of having meningitis should be evaluated with lumbar puncture and the cerebrospinal fluid (CSF) should be sent for routine cell count, protein and glucose concentrations, as well as smears and cultures.

For mild presentations of acute sinusitis or otitis media, clinicians often make the diagnosis on clinical grounds. Plain sinus radiographs may be useful in the initial evaluation and management of patients with suspected sinusitis. If the presentation requires more aggressive evaluation, such as when empiric therapy fails, diagnosis may require imaging studies and/or an antral puncture to obtain organisms for culture. Magnetic resonance scanning and computed tomography are more sensitive than plain film radiographs and may demonstrate posterior paranasal sinus involvement.(42)

As in the general population, the treatment of HIV-1 related pneumococcal disease depends on the site of infection and whether or not the organism is susceptible to penicillin. HIV-1 seropositive patients with moderate to severe pneumococcal pneumonia should be hospitalized and treated with parenteral antimicrobial agents (Table 1). For hospitalized patients with documented penicillin-sensitive (MIC <0.1 micrograms/mL) pneumococcal pneumonia, penicillin (500,000 to 2,000,000 units intravenously every 4 to 6 hours) is the drug of choice. Alternative antibiotics would include ampicillin, first generation cephalosporins, macrolides, and clindamycin. Milder cases of community-acquired pneumococcal pneumonia may be treated with an oral regimen as an outpatient: oral penicillins, first or second generation cephalosporins, or macrolides would be appropriate (see Table 1). Most patients should be treated for approximately 10 days. The duration of therapy is the same regardless of whether or not the patient received intravenous or oral regimens.

In patients with atypical presentations (eg, diffuse infiltrates, subacute clinical course) or a nondiagnostic Gram's stain, it is reasonable to begin treatment for PCP with TMP-SMX (20 mg trimethoprim/kg) because this agent also provides coverage for most H. influenzae and S. pneumoniae isolates in adults. If possible, clinicians should avoid TMP-SMX for the treatment of known bacterial pneumonia because of the high incidence of serious side effects associated with its use in HIV-1 infected patients.(51) If the diagnosis is unclear or if the patient cannot tolerate TMP-SMX, clinicians should start treatment with pentamidine and an antimicrobial agent effective against encapsulated bacteria while diagnostic tests for P. carinii are pending.

Patients with suspected pneumococcal meningitis should be treated with an extended spectrum cephalosporin, such as ceftriaxone. Documented penicillin-sensitive pneumococcal meningitis should be treated with 12 to 24 million units of penicillin intravenously daily given in divided doses. Patients who are allergic to penicillin can be treated with an extended spectrum cephalosporin.

Sinusitis and otitis media are different from the infections just described because they are usually treated empirically with oral antibiotics. In the setting of HIV-1 infection, most authorities recommend standard therapy, such as amoxicillin-clavulanate or an oral cephalosporin.(41) The duration of therapy is 10 days. In cases of chronic sinusitis or when empiric therapy fails, broader coverage is needed because of an increased frequency of Pseudomonas aeruginosa. Additionally, a surgical approach should be considered. Systemic decongestants and guaifenesin may be useful adjuncts to antibiotic therapy.(41)

Pulmonary infections caused by penicillin-resistant pneumococci can be treated with higher doses of penicillin (2 million units intravenously every 2 to 4 hours), as long as the level of resistance is low to intermediate (MIC 0.1 to 1.0 g/mL). In patients with high level resistance (MIC >=2 g/mL), vancomycin, a fluoroquinolone, or imipenem/cilastatin is recommended.(52) Extended spectrum cephalosporins, such as cefotaxime or ceftriaxone, are the preferred initial empiric therapy for suspected pneumococcal meningitis. Because there have been treatment failures with cephalosporin therapy, combination therapy with ceftriaxone and vancomycin, with or without rifampin, has been proposed.(53) Clindamycin has excellent activity against penicillin-resistant organisms and has been used successfully (with or without rifampin) in children with AOM who failed beta-lactam antibiotics.(52)

Despite the presence of underlying immunodeficiency and a high frequency of bacteremia, HIV-1 infected patients generally respond rapidly to antimicrobial therapy and no extended therapy is needed. Patients with pneumococcal pneumonia who fail to respond to appropriate antimicrobial therapy or whose condition worsens after initial improvement may have a coexistent opportunistic infection, such as PCP.(Figure 3)

The mortality rate for invasive pneumococcal disease has been relatively low, with no deaths reported in several studies.(3,26,33,49,54) In a retrospective review of invasive pneumococcal disease in 147 hospitalized patients, Frankel and colleagues(5) reported an overall mortality of 12%; the mortality rate did not differ by HIV-1 serostatus.

Despite the good response to therapy, recurrent bacterial disease is more common in patients with HIV-1 infection. Recurrent pneumococcal infection has been reported in 8% to 25% of HIV-1 infected patients compared with 7% of controls.(3,5,6,26) Using serotyping, ribotyping, and DNA typing, investigators have demonstrated that recurrences can occur because of relapse with the same strain or reinfection with another strain of S. pneumoniae.(55) Recurrent bacterial pneumonia is now a reporting criteria for AIDS.(56)

The increased incidence of pneumococcal infections and the high rate of bacteremia and recurrence make prevention of invasive pneumococcal disease an important aspect of the care of HIV-1 infected patients. Possible preventive strategies include pneumococcal vaccination, intravenous immunoglobulin, and oral antibiotics. The Advisory Committee on Immunization Practices of the CDC recommends that all HIV-1 infected persons over 2 years of age receive the 23-valent pneumococcal polysaccharide vaccine.(57) In a recent CDC study,(58) however, only 37% of a cohort of 9,737 HIV-1 infected persons 13 years of age or older had received the vaccine. Compliance with the CDC recommendation may be poor because the efficacy of the pneumococcal vaccine in HIV-1 infected patients has not been established.

HIV-1 infection decreases the antibody response to capsular polysaccharide vaccines; the more severe the underlying immunosuppression, the lower the antibody titers.(25,28,59,60) Some individuals, however, are able to mount a protective antibody response to vaccination.(60-63) One study measured the total IgG antibody response against all 23 S. pneumoniae capsular antigens. Eighty-three percent of the subjects, all HIV-1 seropositive, responded to the vaccine; the ability to respond to vaccination did not correlate with the CD4 lymphocyte count. The antibody response to pneumococcal vaccine is significantly higher for persons with symptomatic and advanced HIV-1 disease who are receiving zidovudine therapy, than for those who are not taking zidovudine.(64) The effect of protease inhibitors on the antibody response to pneumococcal vaccine is not known.

The CDC currently recommends that HIV-1 seropositive individuals receive a single revaccination 5 years or more after receipt of the first dose.(57) Whether or not HIV-1 infected patients will benefit from revaccination is unclear. In HIV-1 infected subjects who were unable to respond to vaccination with a 23-valent vaccine, revaccination with a double dose did not stimulate an IgG response.(65) On the other hand, individuals who responded initially to vaccination maintained geometric mean IgG levels similar to HIV-1 seronegative subjects at 1 and 2 years of follow-up.

Conjugation of T-cell-independent polysaccharide antigens to proteins is thought to elicit a T-cell-dependent immune response.(60) Haemophilus influenzae type b conjugate vaccine has been shown to be more immunogenic in HIV-1 infected patients than the standard Hib polysaccharide vaccine,(66) although a conjugated pneumococcal vaccine was reported to be of similar immunogenicity as that of the polysaccharide vaccine in HIV-1 infected patients.(60) Among HIV-1 seronegative patients, the conjugate vaccine was superior.

A 1996 report described an increase in HIV-1 replication in asymptomatic seropositive subjects who received a pneumococcal vaccine.(67) The magnitude of these increases correlated with the extent of the antibody response to the vaccination. Another study, however, did not demonstrate any increase in viral load 6 weeks after vaccination.(68) A transient increase in viral load has been demonstrated after influenzae(69,70) and tetanus toxoid vaccination.(71) Whether this phenomenon has any clinical significance and whether or not protease inhibitors will block this increase in viral replication is not known.

Studies addressing the efficacy of intravenous immunoglobulin G (IVG) have been reviewed extensively by Yap.(72) IVG decreases the rate of both minor and serious bacterial infections in HIV-1 infected children with entry CD4 count of at least 200 cells/mm³.(14,73) The efficacy of gammaglobulin in preventing (or treating) pneumococcal infections in HIV-infected adults is still unclear. Although one study demonstrated a decrease in mortality in HIV-infected patients treated with intravenous immunoglobulin (IVG), no significant reduction was noted in the rate of bacterial infections.(74) The only placebo-controlled trial of IVG conducted in adults reported no benefit in IVG over placebo in terms of mortality or rate of bacterial infections.(75) Two other studies evaluated mortality and the rate of infection in patients with AIDS, and in both trials, a significant reduction in mortality was observed, although neither trial was double-blind nor placebo-controlled.(72) At this time, routine use of IVG in adults with AIDS cannot be recommended, although further studies are necessary in patients with recurrent bacterial infections. In HIV-1-infected children, however, IVG is currently recommended for the treatment of recurrent bacterial infections.(72)

A 1992 report indicated that oral TMP-SMX significantly decreased the risk of bacterial infection in adults with AIDS.(76) In a large multicenter trial comparing TMP-SMX with aerosolized pentamidine for PCP prophylaxis, the incidences of bacterial diseases were lower and the time to first bacterial infection longer in those who were taking the oral antibiotic.(76) In addition, oral TMP-SMX used as PCP prophylaxis has been associated with a reduction in the episodes of bacterial pneumonia(11) and a lower risk of community-acquired bacteremia. Whether or not this approach will decrease pneumococcal infections is not clear. Some clinicians offer antibiotic prophylaxis to patients who have experienced recurrent invasive pneumococcal infections. Given the rising resistance of S. pneumoniae to TMP-SMX and penicillin, this approach cannot be recommended routinely.